Happy Halloween! Today, Jon Foust (https://twitter.com/syntxerror1) and Brian Dorsey (https://twitter.com/briandorsey) chat with Maria Laura Scuri (https://www.linkedin.com/in/maria-laura-scuri-43270689/) of FACEIT about ways they are reducing toxicity in gaming. FACEIT is a competitive gaming platform that helps connect gamers and game competition and tournament organizers. In order to do this well, FACEIT has put a lot of energy into finding ways to keep the experience positive for everyone.

Because gaming toxicity can involve anything from verbal jabs to throwing a game, FACEIT uses a combination of data collecting programs and input from players to help identify toxic behavior. In identifying this behavior, FACEIT has to consider not only the literal words spoken or actions made, but the context around them. Is that player being rude to strangers or is he egging on a friend? The answer to this question could change the behavior from unacceptable to friendly banter. Using their own machine learning model, interactions are then given a score to determine how toxic the player was in that match.

The toxicity scores along with their program, Minerva, determine if any bans should be put on a player. FACEIT focuses on punishing player behavior, rather than the player themselves, in an effort to help players learn from the experience and change the way they interact with others in the future.

Maria’s advice to other companies looking to help reduce toxicity on their platforms is to know the context of the toxic event. Know how toxicity can express itself on your platform and find ways to deal with all of them. She also suggests tackling the issues of toxicity in small portions and celebrating the small wins! Her final piece of advice is to focus on criticizing the behavior of the user rather than attacking them personally.

Maria Laura Scuri

Maria (https://www.linkedin.com/in/maria-laura-scuri-43270689/) is the Director of Business Intelligence at FACEIT, the leading competitive platform for online multiplayer games with over 15 million users. She joined FACEIT as part of the core team in 2013 as an intern assisting with everything from customer support to event management. Her passion for data, machine learning, and artificial intelligence saw her quickly rise through the ranks to her current position, leading the Business Intelligence and Data Science teams.

Maria works side by side with some of the biggest tech companies in the world including Google Cloud. She is the main lead on a number of projects including the inception of an Artificial Intelligence Admin to fight toxicity on the platform. Maria is responsible for implementing best practices around data visualization and tools that allow the FACEIT team to thrive, as well as sourcing and training new talent.

Maria is a huge video games fan. You can find her on League of Legends as “FACEIT Lulu” and on Steam as “Sephariel (https://steamcommunity.com/id/sephariel89/)”.

Cool things of the week

* What can Google Cloud do for you? New trainings for business professionals blog (https://cloud.google.com/blog/topics/training-certifications/new-google-cloud-trainings-for-business-professionals)

* Leave no database behind with Cloud SQL for SQL Server blog (https://cloud.google.com/blog/products/databases/leave-no-database-behind-with-cloud-sql-for-sql-server)

* How to orchestrate Cloud Dataprep jobs using Cloud Composer blog (https://cloud.google.com/blog/products/data-analytics/how-to-orchestrate-cloud-dataprep-jobs-using-cloud-composer)

* Updates make Cloud AI platform faster and more flexible blog (https://cloud.google.com/blog/products/ai-machine-learning/updates-make-cloud-ai-platform-faster-and-more-flexible)

* Use GKE usage metering to combat over-provisioning blog (https://cloud.google.com/blog/products/containers-kubernetes/use-gke-usage-metering-to-combat-over-provisioning)

Interview

* FACEIT site (https://www.faceit.com/en)

We’re sad to say goodbye to Mark Mandel (https://twitter.com/Neurotic) this week but excited to bring you an interview he and guest host Robert Martin (https://twitter.com/cloudrobx) did with Björn Lindberg of Massive Entertainment (https://www.massive.se/). The gaming studio is located in Sweden and owned by Ubisoft. Their most recent game, The Division 2 (https://tomclancy-thedivision.ubisoft.com/game/en-us/home), is a “looter shooter” game that was released in March. It can be played solo or users can be matched up to play with or against others.

To keep the game running smoothly, Massive employs a micro-service architecture to divide and conquer the trials of creating and running such a large, intense game. The Division 2 was launched with Google Cloud, a process Björn says was a bit easier than launching on physical hardware. Autoscaling in the cloud has created a simpler, more trustworthy gaming process as well, and by connecting to data centers in multiple regions, they’re able to decrease latency.

Björn Lindberg

Björn Lindberg is working as On-Line technical director at Massive Entertainment a Ubisoft owned and operated game studio in Malmö Sweden. He does design and implementation of on-line backend systems for large AAA on-line games such as The Division series of games and World in Conflict.

Interview

* Massive Entertainment site (https://www.massive.se/)

* The Division 2 site (https://tomclancy-thedivision.ubisoft.com/game/en-us/home)

* Ubisoft site (https://www.ubisoft.com)

* Terraform site (https://www.terraform.io)

* Grafana site (https://grafana.com)

* Compute Engine site (https://cloud.google.com/compute/)

Thank You Mark!

Thank you Mark for everything you’ve done to make this podcast a success! We’ll miss you!(http://feeds.feedburner.com/~r/GcpPodcast/~4/iXvoXSFrObQ)

Gabi Ferrara (https://twitter.com/gabidavila) and Jon Foust (https://twitter.com/syntxerror1) are back today and joined by fellow Googler Manuel Lima (https://twitter.com/mslima). In this episode, Manuel tells us all about data visualization, what it means, why it’s important, and the best ways to do it effectively.

For Google and its mission, data visualization is especially necessary in faciliatating the accesibility of information. It “makes the invisible visible” because of the way it can decode meaningful data patterns. Working across multiple GCP products, Manuel and his team build advanced visualization models that go beyond graphs and bar charts to things like sophisticated time lines that aid in the progression from data to usable knowledge. They have also created guidelines for things like what kind of graphical language to use, what type of charts users might need, and more. These guidelines, originally used only internally, have now been adjusted and released for use by developers outside Google with the help of the Material.io team.

The guidelines are based around the six data visualazation princples that help users get started. They can be employed to plan and inspire an entire project or to evaluate a specific data visualation chart. Some of the most important principles are to be honest and to lend a helping hand. You can read more in their Medium article, Six Principles for Designing Any Chart (https://medium.com/google-design/redefining-data-visualization-at-google-9bdcf2e447c6).

Manuel Lima

A Fellow of the Royal Society of Arts and nominated by Creativity magazine as “one of the 50 most creative and influential minds of 2009,” Manuel Lima (https://twitter.com/mslima) is the founder of VisualComplexity.com, Design Lead at Google, and a regular teacher of data visualization at Parsons School of Design.

Manuel is a leading voice on information visualization and has spoken at numerous conferences, universities, and festivals around the world, including TED, Lift, OFFF, Eyeo, Ars Electronica, IxDA Interaction, Harvard, Yale, MIT, Columbia, the Royal College of Art, NYU Tisch School of the Arts, ENSAD Paris, the University of Amsterdam, and MediaLab-Prado Madrid. He has also been featured in various publications and media outlets, such as Wired, the New York Times, Science, Nature, Businessweek, Fast Company, Forbes, The Guardian, BBC, CNN, Design Observer, Creative Review, Eye, Grafik, étapes, and El País. His first book, Visual Complexity: Mapping Patterns of Information, has been translated into French, Chinese, and Japanese. His latest, The Book of Circles: Visualizing Spheres of Knowledge, covers 1,000 hundred years of humanity’s long-lasting obsession with all things circular.

With more than twelve years of experience designing digital products, Manuel has worked for Codecademy, Microsoft, Nokia, R/GA, and Kontrapunkt. He holds a BFA in Industrial Design and a MFA in Design & Technology from Parsons School of Design. During the course of his MFA program, Manuel worked for Siemens Corporate Research Center, the American Museum of Moving Image, and Parsons Institute for Information Mapping in research projects for the National Geospatial-Intelligence Agency.

Cool things of the week

* Compute Engine or Kubernetes Engine? New trainings teach you the basics of architecting on Google Cloud blog (https://cloud.google.com/blog/topics/training-certifications/new-trainings-teach-you-the-basics-of-architecting-on-google-cloud)

* Stadia comes next month site (https://store.google.com/us/product/stadia_learn?hl=en-US)

* Google Cloud named a Leader in the 2019 Gartner Magic Quadrant for Full Life Cycle API Management for the fourth consecutive time blog (https://cloud.google.com/blog/products/apigee/google-cloud-named-a-leader-in-2019-gartner-magic-quadrant-full-life-cycle-api-management-for-fourth-consecutive-time)

* Google Hardware Event

* Pixel 4 is here to help blog (https://www.blog.

Today on the podcast, Gabi Ferrara (https://twitter.com/gabidavila) and Jon Foust (https://twitter.com/syntxerror1) share a great interview with Laura Ham (https://twitter.com/laura_hamham), Community Solution Engineer at SeMI Technologies (https://twitter.com/SeMI_tech). At SeMI Technologies, Laura works with their project Weaviate, an open-source knowledge graph program that allows users to do a contextualized search based on inputted data. However, unlike traditional databases, Weaviate attaches meanings and links within the data.

Laura details what knowledge graphs are and how they can be useful for both small and large projects. Explaining that ontology is the meaning of words, she tells us how Weaviate is able to use this concept to make more specific data entries and links, allowing users to perform better and more informative searches. Weaviate is able to do this with the help of Kubernetes. Later, Laura tells Gabi and Jon the ways Weaviate helps developers and users with thorough documentation, assistance with troubleshooting, and support from solution engineers.

Laura Ham

Laura (mailto: [email protected]) is the Community Solution Engineer at SeMI Technologies, where she takes care of building and supporting a community around their open source software product, Weaviate. She also takes care of the developer and user experience within the business, which means she writes documentation to support both developers and users, as well as researches and evaluates new software implementations on user experience. She has a user-centered approach in the work that she develops and designs. Laura is a full-time graduate student in Human Computer Interaction and Design with a special focus on Innovation and Entrepreneurship at EIT Digital Master School. Here, she learns about how to develop and design technology from a user perspective, and how to apply this with an entrepreneurial mindset.

Cool things of the week

* Use G Suite to make documents (and other tools) more accessible to people with disabilities blog (https://cloud.google.com/blog/products/g-suite/use-g-suite-to-make-documents-and-other-tools-more-accessible-to-people-with-disabilities)

* 4 steps to stop data exfiltration with Google Cloud blog (https://cloud.google.com/blog/products/identity-security/4-steps-to-stop-data-exfiltration-with-google-cloud)

* Using Colab to get more out of BigQuery blog (https://cloud.google.com/blog/products/data-analytics/whats-the-weather-like-using-colab-to-get-more-out-of-bigquery)

* Updates to AutoML Vision Edge, Auto ML Video, & AutoML Intelligence API blog (https://cloud.google.com/blog/products/ai-machine-learning/announcing-updates-to-automl-vision-edge-automl-video-and-video-intelligence-api)

Interview

* SeMI Technologies site (https://www.semi.technology)

* Weaviate site (https://www.semi.technology/products/weaviate.html)

* Weaviate GitHub github (https://github.com/semi-technologies/weaviate)

* Weaviate documentation site (https://www.semi.technology/documentation/weaviate/current/)

* GraphQL API github (https://github.com/semi-technologies/weaviate-graphql-prototype)

* Kubernetes site (https://kubernetes.io)

* GKE site (https://cloud.google.com/kubernetes-engine/)

* Cloud BigTable site (https://cloud.google.com/bigtable/)

* SeMI Technologies Meetups site (https://www.semi.technology/news/meetups.html)

Question of the week

When will Python 2 reach the end of its life, and what does that mean for GCP?

* Python Google Cloud Client Libraries only support Python 3 github (https://github.com/googleapis/google-cloud-python)

* Countdown to end of life site (https://pythonclock.org/)

Where can you find us next?

Gabi will be at Full Stack Europe (https://fullstackeurope.com/speakers/gabi-davila/).

Jon will be at his twin’s wedding! Then Kubecon (https://events19.linuxfoundation.org/events/kubecon-cloudnativecon-north-america-2019/) in November to run a workshop on Open Match (https://github.

Our guests Matthew Tamsett and Ravi Upreti join Gabi Ferrara (https://twitter.com/gabidavila) and Aja Hammerly (https://twitter.com/the_thagomizer) to talk about data science and their project, Qubit. Qubit helps web companies by measuring different us...

Mark Mandel (https://twitter.com/Neurotic) and Jon Foust (https://twitter.com/syntxerror1) return this week to host Jesse Houston (https://twitter.com/gtez), CEO of Phoenix Labs. Jesse goes into detail about their online, multiplayer game Dauntless, a hunting action game that brings friends together from every platform to fight giant monsters. Users can even switch platforms, say from Xbox to Playstation, and pick up right where they left off.

Later in the show, Jesse describes the hurdles of building such a huge game and how Phoenix Labs overcame them. Late nights and holiday hours helped them create “no downtime deploys”, so users can continue to play even as the game updates. Because big projects sometimes come with big problems, Jesse also emphasized the importance of developing crisis management skills to help get through tough times. We talk more specifically about what it takes to build and run Dauntless, from GCP products such as GKE, Bigtable, and BigQuery, to tricks with scaling and management. In the future, Dauntless will be available on the Switch, new expansions will be released, and more.

Jesse Houston

Jesse Houston (https://twitter.com/gtez) is a games industry veteran with over 18 years experience in the gaming space. Houston fell in love with games at an early age and found his footing in the games industry by applying to a QA position in a local paper. Previously, Houston has held lead producer roles at both Riot Games on League of Legends, and BioWare on the Mass Effect series. He also served as Production Director at Ubisoft, overseeing Technical Project Management, Pipeline Planning, Development and Design, among other responsibilities. Houston formed Phoenix Labs with Sean Bender and Robin Mayne to create deep multiplayer games that bring players together.

Cool things of the week

* Virtual display devices for Compute Engine are now GA blog (https://cloud.google.com/blog/products/compute/virtual-display-devices-for-compute-engine-now-ga)

* Container-native load balancing on GKE are now GA blog (https://cloud.google.com/blog/products/containers-kubernetes/container-native-load-balancing-on-gke-now-generally-available)

* How to deploy a Windows container on Google Compute Engine blog (https://cloud.google.com/blog/products/containers-kubernetes/how-to-deploy-a-windows-container-on-google-compute-engine)

* Agones 1.0 site (https://agones.dev/site/blog/2019/09/16/1.0.0-a-huge-milestone/)

Interview

* Phoenix Labs site (https://phxlabs.ca)

* Dauntless site (https://playdauntless.com/landing/)

* Dauntless Updates site (https://playdauntless.com/roadmap/)

* Dauntless on Twitter twitter (https://twitter.com/PlayDauntless)

* Cloud Bigtable site (https://cloud.google.com/bigtable/)

* Kubernetes site (https://kubernetes.io)

* BigQuery site (https://cloud.google.com/bigquery/)

* GKE site (https://cloud.google.com/kubernetes-engine/)

* Redis site (https://redis.io)

* Introducing Google Customer Reliability Engineering blog (https://cloud.google.com/blog/products/gcp/introducing-a-new-era-of-customer-support-google-customer-reliability-engineering)

* Cloud SQL site (https://cloud.google.com/sql/)

Question of the week

What is the difference between Premium vs Standard network? (https://cloud.google.com/network-tiers/)

Where can you find us next?

Jesse will be speaking at the Montreal International Game Summit (https://megamigs.com/en/migs-home/). You can see Phoenix Labs at many other gaming conferences, including Pax (https://www.paxsite.com), TwitchCon (https://www.twitchcon.com), and GDC (https://www.gdconf.com).

Mark is taking some vacation time, then he’ll be at Kubecon (https://events.linuxfoundation.org/events/kubecon-cloudnativecon-north-america-2019/).

Jon will also be at Kubecon (https://events.linuxfoundation.org/events/kubecon-cloudnativecon-north-america-2019/), as well as taking some personal time to attend several weddings.

Sound Effect Attribution

Conversational AI is our topic this week as your hosts Mark Mirchandani (https://twitter.com/markmirch) and Priyanka Vergadia (https://twitter.com/pvergadia) are joined by Cathy Pearl (https://twitter.com/cpearl42) and Jessica Dene Earley-Cha (https://twitter.com/chatasweetie). Cathy explains what conversation AI is, describing it as people teaching computers to communicate the way humans do, rather than forcing humans to communicate like computers.

Later, we talk best practices in design and development, including how a good conversation design and sample dialogues before building can create a better product. This prep work helps anticipate the ways different users could respond to the same question and how the program should react. In multi-modal programming, planning is also important. Our guests suggest starting with the spoken portions of the design and then planning visual components that would augment the experience. Working together as a team is one of the most important parts of the planning process.

We also talk best use-cases for conversation AI. Does performing this task via voice make the experience better? Does it make the task easier or more accessible? If so, that could be a great application. In the future, the conversation may be a silent communication with the help of MIT’s Alter Ego.

Cathy Pearl

Cathy Pearl (https://twitter.com/cpearl42) is head of conversation design outreach and author of the O’Reilly book, “Designing Voice User Interfaces”. She’s been creating Voice User Interfaces for 20 years and has worked on everything from programming NASA helicopter pilot simulators to a conversational app in which Esquire’s style columnist advises what to wear on a first date. She earned an MS in Computer Science from Indiana University and a BS in Cognitive Science from UC San Diego. You can find Cathy on Twitter, or check out her latest Medium article “A Conversation With My 35-year-old Chatbot” (https://chatbotslife.com/a-conversation-with-my-35-year-old-chatbot-part-1-21f0b3ecfcc).

Jessica Dene Earley-Cha

Jessica Dene Earley-Cha (https://twitter.com/chatasweetie) is a Developer Advocate for Actions on Google. She loves to connect with developers and explore VUI (voice user interface) to add another dimension to how users interact with technology. Jessica is part of the leadership team for @WomenInVoice. You’ll find her either spending time with her dog, collecting strawberry knick knack or biking around town. Stay up-to-date on her ventures on Twitter.

Cool things of the week

* How Google and Mayo Clinic will transform the future of healthcare blog (https://cloud.google.com/blog/topics/customers/how-google-and-mayo-clinic-will-transform-the-future-of-healthcare)

* Announcing the general availability of 6 and 12 TB VMs for SAP HANA instances on Google Cloud Platform blog (https://cloud.google.com/blog/products/sap-google-cloud/announcing-the-general-availability-of-6-and-12tb-vms-for-sap-hana-instances-on-gcp)

* Understanding your GCP Costs site (https://www.qwiklabs.com/quests/90) and videos (https://www.youtube.com/playlist?list=PLIivdWyY5sqKJx6FwJMRcsnFIkkNFtsX9)

* Coupon code for qwiklabs is: 1q-costs-626

Interview

* GCP Podcast Episode 188: Conversation AI with Priyanka Vergadia podcast (https://www.gcppodcast.com/post/episode-188-conversation-ai-with-priyanka-vergadia/)

* Google’s Conversation Design Best Practices site (https://designguidelines.withgoogle.com/conversation/)

* Actions on Google site (https://developers.google.com/actions/)

* Interactive Canvas docs (https://developers.google.com/actions/interactivecanvas/)

* Dialogflow site (https://dialogflow.com)

* Deconstructing Chatbots videos (https://www.youtube.com/playlist?list=PLIivdWyY5sqK5SM34zbkitWLOV-b3V40B)

* Behind the Actions videos (https://www.youtube.com/playlist?list=PLOU2XLYxmsIKP4Hh9gQO54naZ8V7mDEQi)

* Assistant On Air videos (https://www.youtube.com/playlist?

On the podcast this week, we have a great interview with Google Developer Advocate, Dale Markowitz (https://twitter.com/dalequark). Aja Hammerly (https://twitter.com/the_thagomizer) and Jon Foust (https://twitter.com/syntxerror1) are your hosts, as we talk about machine learning, its best use cases, and how developers can break into machine learning and data science. Dale talks about natural language processing as well, explaining that it’s basically the intersection of machine learning and text processing. It can be used for anything from aggregating and sorting Twitter posts about your company to sentiment analysis.

For developers looking to enter the machine learning space, Dale suggests starting with non life-threatening applications, such as labeling pictures. Next, consider the possible mistakes the application can make ahead of time to help mitigate issues. To help prevent the introduction of bias into the model, Dale suggests introducing it to as many different types of project-appropriate data sets as possible. It’s also important to continually monitor your model.

Later in the show, we talk Google shop, learning about all the new features in Google Translate and AutoML.

Dale Markowitz

Dale Markowitz (https://twitter.com/dalequark) is an Applied AI Engineer and Developer Advocate for ML on Google Cloud. Before that she was a software engineer in Google Research and an engineer at the online dating site OkCupid.

Cool things of the week

* Build a dev workflow with Cloud Code on a Pixelbook blog (https://cloud.google.com/blog/products/application-development/build-a-dev-workflow-with-cloud-code-on-a-pixelbook)

* Feminism & Agile blog (https://medium.com/@Hanna.Thomas/why-dont-we-just-call-agile-what-it-is-feminist-8bdd9193edba)

* New homepage and improved collaboration features for AI Hub blog (https://cloud.google.com/blog/products/ai-machine-learning/new-homepage-and-improved-collaboration-features-for-ai-hub)

Interview

* TensorFlow site (https://www.tensorflow.org)

* Natural Language API site (https://cloud.google.com/natural-language/)

* AutoML Natural Language site (https://cloud.google.com/natural-language/#how-automl-natural-language-works)

* Content Classification site (https://cloud.google.com/natural-language/docs/classifying-text)

* Sentiment Analysis site (https://cloud.google.com/natural-language/docs/analyzing-sentiment)

* Analyzing Entities site (https://cloud.google.com/natural-language/docs/analyzing-entities)

* Translation API site (https://cloud.google.com/translate/)

* AutoML Translate site (https://cloud.google.com/translate/#automl-translation)

* Google Translate Glossary Documentation docs (https://cloud.google.com/translate/docs/glossary)

* Google News Lab site (https://newsinitiative.withgoogle.com/google-news-lab)

* AI Platform’s Data Labeling Service docs (https://cloud.google.com/data-labeling/docs/)

Question of the week

How many different ways can you run a container on GCP?

* GKE (https://cloud.google.com/kubernetes-engine/)

* Cloud Run (https://cloud.google.com/run/)

* App Engine Flexible Environmnet (https://cloud.google.com/appengine/docs/flexible/)

* Compute Engine (https://cloud.google.com/compute/)

* VM as a computer

Where can you find us next?

Dale will be at DevFest Minneapolis (https://devfest.mn), DevFest Madison (https://devfestwi.com), and London NEXT (https://cloud.withgoogle.com/next/uk).

Jon will be at the internal Google Game Summit and visiting Montreal.

Aja will be holding down the fort at home.

Sound Effect Attribution

* “Mystery Peak2” by FoolBoyMedia of Freesound.org (https://freesound.org)

* “Collect Point 00” by LittleRobotSoundFactory of Freesound.org (https://freesound.org)

* “Cinematic Piano” by Ellary of Freesound.org (https://freesound.org)

(http://feeds.feedburner.com/~r/GcpPodcast/~4/M5jOlpc17YQ)

Michelle Casbon (https://twitter.com/texasmichelle) is back in the host seat with Mark Mirchandani (https://twitter.com/markmirch) this week as we talk data science with Devoted Health Director of Data Science, Chris Albon (https://twitter.com/chrisalbon). Chris talks with us about what it takes to be a data scientist at Devoted Health and how Devoted Health and machine learning are advancing the healthcare field. Later, Chris talks about the future of Devoted Health and how they plan to grow. They’re hiring!

At Devoted Health, they emphasize knowledge, supporting a culture of not just machine learning but people learning as well. Questions are encouraged and assumptions are discouraged in a field where a tiny mistake can change the care a person receives. Because of this, their team members not only have a strong data science background, they also learn the specific nuances of the healthcare system in America, combined with knowledge of the legal and privacy regulations in that space.

How did Chris go from Political Science Ph.D. to non-profit data science wizard? Listen in to find out his storied past.

Chris Albon

Chris Albon (https://twitter.com/chrisalbon) is the Director of Data Science at Devoted Health, using data science and machine learning to help fix America’s health care system. Previously, he was Chief Data Scientist at the Kenyan startup BRCK, cofounded the anti-fake news company New Knowledge, created the data science podcast Partially Derivative, led the data team at the humanitarian non-profit Ushahidi’s, and was the director of the low-resource technology governance project at FrontlineSMS. Chris also wrote Machine Learning For Python Cookbook (O’Reilly 2018) and created Machine Learning Flashcards.

He earned a Ph.D. in Political Science from the University of California, Davis researching the quantitative impact of civil wars on health care systems. Chris earned a B.A. from the University of Miami, where he triple majored in political science, international studies, and religious studies.

Cool things of the week

* How Itaú Unibanco built a CI/CD pipeline for ML using Kubeflow blog (https://cloud.google.com/blog/products/ai-machine-learning/itau-unibanco-how-we-built-a-cicd-pipeline-for-machine-learning-with-online-training-in-kubeflow)

* Why TPUs are so high-performance

* BFloat16: The secret to high performance on Cloud TPUs blog (https://cloud.google.com/blog/products/ai-machine-learning/bfloat16-the-secret-to-high-performance-on-cloud-tpus)

* TPU Codelabs site (https://codelabs.developers.google.com/codelabs/keras-flowers-data/#0)

* Benchmarking TPU, GPU, and CPU Platforms for Deep Learning paper (https://arxiv.org/abs/1907.10701)

* Machine Learning Flashcards site (https://machinelearningflashcards.com)

Interview

* Devoted Health site (https://www.devoted.com)

* Devoted Health is hiring! site (https://jobs.lever.co/devoted/d0758ba1-3bde-42c6-9981-b28f2041e461)

* Ushahidi site (https://www.ushahidi.com)

* FrontlineSMS site (http://www.frontlinesms.com)

* New Knowledge site (http://www.newknowledge.io)

* Joel Grus: Fizz Buzz in TensorFlow site (https://joelgrus.com/2016/05/23/fizz-buzz-in-tensorflow/)

* Snowflake site (https://www.snowflake.com)

* Periscope Data site (https://www.periscopedata.com)

* Airflow site (https://airflow.apache.org/#)

* Kubernetes site (https://kubernetes.io)

* Chris Albon’s Website site (https://chrisalbon.com)

* Partially Derivative podcast (http://partiallyderivative.com)

* Partially Derivative Back Episodes podcast (http://partiallyderivative.com/podcast/)

Question of the week

(../post/episode-193-devoted-health-and-data-science-with-chris-albon/images/learn-in-machine-learning.png)

Chris Albon

To paraphrase: A computer program is said to learn if its performance at specific tasks improves with experience.

To find out more, including the definition of a partial derivative, buy a pack of Chris’s flashcards. Who knows,

Google’s own Billy Jacobson (https://twitter.com/billyjacobson) joins hosts Mark Mandel (https://twitter.com/Neurotic) and Mark Mirchandani (https://twitter.com/markmirch) this week to dive deeper into Cloud Bigtable. Bigtable is Google’s petabyte scale, fully managed, NoSQL database. Billy elaborates on what projects Bigtable works best with, like time-series data user analytics, and why it’s such a great tool. It offers huge scalability with the benefits of a managed system, and it’s flexible and easily customized so users can turn on and off the pieces they need.

Later, we learn about other programs that are compatible with Bigtable, such as JanusGraph, Open TSDB, and GeoMesa. Bigtable also supports the API for HBase, an open-source project similar to Bigtable. Because of this, it’s easy for HBase users to move to Bigtable, and the Bigtable community has access to many open source libraries. Billy also talks more about the nine clients available, and when customers might want to use Bigtable instead of, or in conjunction with, other Google services such as Spanner and BigQuery.

Billy Jacobson

Billy Jacobson (https://twitter.com/billyjacobson) is a developer programs engineer focusing on Cloud Bigtable.

Cool things of the week

* Introducing Cloud Run Button: Click-to-deploy your git repos to Google Cloud blog (https://cloud.google.com/blog/products/serverless/introducing-cloud-run-button-click-to-deploy-your-git-repos-to-google-cloud)

* Firebase Unity Solutions: Update game behavior without deploying with Remote Config blog (https://cloud.google.com/blog/products/gaming/firebase-unity-solutions-update-game-behavior-without-deploying-with-remote-config)

* Introducing the BigQuery Terraform module blog (https://cloud.google.com/blog/products/data-analytics/introducing-the-bigquery-terraform-module)

* Macy’s uses Google Cloud to streamline retail operations blog (https://cloud.google.com/blog/topics/customers/macys-uses-google-cloud-to-streamline-retail-operations)

Interview

* Cloud Bigtable site (https://cloud.google.com/bigtable/)

* GCP Podcast Episode 18: Bigtable with Ian Lewis podcast (https://www.gcppodcast.com/post/episode-18-bigtable-with-ian-lewis/)

* BigQuery site (https://cloud.google.com/bigquery/)

* Bigtable Documentation docs (https://cloud.google.com/bigtable/docs/)

* Codelab: Introduction to Cloud Bigtable site (https://codelabs.developers.google.com/codelabs/cloud-bigtable-intro-java/index.html#0)

* Key Visualizer docs (https://cloud.google.com/bigtable/docs/keyvis-overview)

* Bigtable Replication Documentation docs (https://cloud.google.com/bigtable/docs/replication-overview)

* Bigtable and HBase Documentation docs (https://cloud.google.com/bigtable/docs/hbase-bigtable)

* HBase site (https://hbase.apache.org)

* JanusGraph site (https://janusgraph.org)

* Open TSDB site (http://opentsdb.net)

* GeoMesa site (https://www.geomesa.org)

* Bigtable Client Libraries docs (https://cloud.google.com/bigtable/docs/reference/libraries)

* Cloud Spanner site (https://cloud.google.com/spanner)

* Managing IoT Storage with Google’s Cloud Platform (Google I/O’19) video (https://www.youtube.com/watch?v=2j1BVZOMqWU)

* Cloud Datastore site (https://cloud.google.com/datastore/)

* Cloud Firestore site (https://firebase.google.com/products/firestore)

* Mapping the invisible: Street View cars add air pollution sensors site (https://sustainability.google/projects/airview/)

* Breathing Easy with Bigtable article (https://medium.com/google-cloud/breathing-easy-with-bigtable-b58eb302cc1a)

Question of the week

If I have an organization, how do I break down my billing data by folder? (https://cloud.google.com/blog/products/data-analytics/analyzing-gcp-costs-using-folders-and-bigquery-billing-export)

Where can you find us next?

Mark Mirch is working around town but will be headed to LA soon.

Mark Mandel will be at Pax Dev (https://dev.paxsite.com), Pax West (https://west.paxsite.com), Kubecon (https://events.